[gmx-users] lincs errors late in simulation

[Justin Lemkul]

On 5/14/13 10:27 AM, Joe Smerdon wrote:
>
> Hi all,
>
> First of all I'm new to md simulations, so if I am missing anything simple, I would be grateful for any pointers, however small.  I am trying to simulate many small molecules on a substrate.  I am restraining my substrate from moving but allowing the molecules to do whatever they like.  My steps are to minimise my system (136 molecules on a 200Ax200A surface of about 4 layers, with 100A vacuum), then to run an NVT simulation without pbcs (though I could use pbcs, I just don't really need to for this).  I am getting LINCS errors late on in the simulations - they don't always cause a crash, but sometimes they do.  I guess it can't be insufficient minimisation if the errors arise after the simulation has been underway for awhile (eg, step 15000, 8ps into a 50ps simulation), but I don't really know what it can be.  My timestep is 0.5 fs.  I am just experimenting at the moment, I intend to simulate for much longer when I get it working.
>
> I guess my question is: are LINCS errors that occur late in a simulation due to poor minimisation, or not?  If not, what other factors should I look at?
>

Failures at 8 ps are not considered late.  There are a whole host of problems 
that could be present, ranging from .mdp settings to topologies.  See standard 
advice for diagnosing at 
http://www.gromacs.org/Documentation/Terminology/Blowing_Up.

-Justin

-- 
========================================

Justin A. Lemkul, Ph.D.
Research Scientist
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================